Flood Inundation Mapping for using Sentinel-1 SAR Data for Assam during 2018

Abstract

This paper is focusing on the flood inundation mapping of Hailakandi, Assam flood during 2018 using Sentinel 1 Single Look Complex (SLC) and Ground Range Detected (GRD) data. For the analysis of water pixels, pre and post event images of SLC and GRD data were downloaded with similar specification such as same date, pat and row. Supervised classification is performed with the pre and post event of the study area with Random Forest classifier. To avoid the misclassification of more features more than 200 training pixels (includes water and non-water samples) from permanent water bodies and other surface features were considered. Improved classification performed with Random Forest Classifier with more number of decision trees. From this study, it is found that SLC data for water pixels classification is giving promising results in comparison with GRD data due to complex phase information of SLC data in extracting more feature information than GRD data.

Keywords: Decision trees, flood, phase information, random forest classification, scattering mechanism

Cite this Article
M. Gomathi, M. Geetha Priya, C. Chandre Gowda1, D. Krishnaveni. Flood Inundation Mapping for using Sentinel-1 SAR Data for Assam during 2018. Research & Reviews: Journal of Space Science & Technology. 2019; 8(2): 16–25p.